Articles

Rituximab maintenance for 2 years in patients with high tumour burden follicular responding to plus (PRIMA): a phase 3, randomised controlled trial

Gilles Salles, John Francis Seymour, Fritz Off ner, Armando López-Guillermo, David Belada, Luc Xerri, Pierre Feugier, Réda Bouabdallah, John Vincent Catalano, Pauline Brice, Dolores Caballero, Corinne Haioun, Lars Moller Pedersen, Alain Delmer, David Simpson, Sirpa Leppa, Pierre Soubeyran, Anton Hagenbeek, Olivier Casasnovas, Tanin Intragumtornchai, Christophe Fermé, Maria Gomes da Silva, Catherine Sebban, Andrew Lister, Jane A Estell, Gustavo Milone, Anne Sonet, Myriam Mendila, Bertrand Coiffi er, Hervé Tilly

Summary Lancet 2010; 377: 42–51 Background Patients with follicular lymphoma can have long survival times, but disease progression typically occurs Published Online 3–5 years after initial treatment. We assessed the potential benefi t of 2 years of rituximab maintenance after fi rst-line December 21, 2010 treatment in patients with follicular lymphoma receiving a rituximab plus . DOI:10.1016/S0140- 6736(10)62175-7 Methods The randomised, open-label PRIMA study was undertaken in 223 centres in 25 countries. 1217 patients with See Comment page 4 previously untreated follicular lymphoma needing systemic therapy received one of three non-randomised Hospices Civils de Lyon, Université Claude Bernard, immunochemotherapy induction regimens used in routine practice. 1019 patients achieving a complete or partial UMR CNRS5239, Pierre-Bénite, response were then randomly assigned to receive 2 years of rituximab maintenance therapy (375 mg/m² every 8 weeks) France (Prof G Salles MD, or observation. Treatment was assigned equally by centralised block randomisation, stratifi ed by induction regimen, Prof B Coiffi er MD); Peter response, region, and centre. Neither the participants nor those giving the interventions, assessing outcomes, and MacCallum Cancer Centre and University of Melbourne, analysing data were masked to group assignments. The primary endpoint was progression-free survival (PFS). Analysis Melbourne, VIC, Australia was by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT00140582. (Prof J F Seymour FRACP); Ghent University, Ghent, Belgium Findings 505 patients were assigned to rituximab maintenance and 513 to observation (one patient died during (Prof F Off ner MD); Hospital Clinic, Barcelona, Spain randomisation). With a median follow-up of 36 months (IQR 30–42), PFS was 74·9% (95% CI 70·9–78·9) in the (A López-Guillermo MD); Charles rituximab maintenance group (130 patients progressed) and 57·6% (53·2–62·0) in the observation group University in Prague, Faculty of (218 progressed; hazard ratio [HR] 0·55, 95% CI 0·44–0·68, p<0·0001). 2 years after randomisation, 361 patients (71·5%) Medicine, University Hospital in the rituximab maintenance group were in complete or unconfi rmed complete response versus 268 (52·2%) in the Hradec Králové, Prague, Czech Republic (D Belada MD); Institut observation group (p=0·0001). Overall survival did not diff er signifi cantly between groups (HR 0·87, 95% CI Paoli Calmettes, Marseille, 0·51–1·47). Grade 3 and 4 adverse events were recorded in 121 patients (24%) in the rituximab maintenance group France (Prof L Xerri MD, and 84 (17%) in the observation group (risk ratio 1·46, 95% CI 1·14–1·87; p=0·0026). Infections (grades 2–4) were the R Bouabdallah MD); CHU de most common adverse event, occurring in 197 (39%) and 123 (24%) patients, respectively (risk ratio 1·62, 95% CI Nancy, Université Henri Poincaré, Nancy, France 1·35–1·96; p<0·0001). (Prof P Feugier MD); Frankston Hospital, Frankston, VIC, Interpretation 2 years of rituximab maintenance therapy after immunochemotherapy as fi rst-line treatment for Australia (J V Catalano FRACP); follicular lymphoma signifi cantly improves PFS. Hôpital Saint-Louis, AP-HP, Paris, France (P Brice MD); Universitario de Salamanca, Funding Groupe d’Etude des Lymphomes de l’Adulte (GELA) and F Hoff mann-La Roche. Salamanca, Spain (D Caballero MD); Hôpital Henri monoclonal rituximab with various chemo- Mondor, AP-HP, Créteil, France Introduction (Prof C Haioun MD); Odense Follicular lymphoma is the second most common therapy regimens can improve patients’ overall survival, University Hospital, Odense, lymphoma subtype and, despite substantial improve- and this combination is now regarded as the standard of Denmark (L M Pedersen MD); ments in survival, disseminated disease is usually care in fi rst-line follicular lymphoma. CHU de Reims, Reims, France incurable.1,2 The disease characteristically responds well Rituximab, in view of its effi cacy, pharmacokinetic (Prof A Delmer MD); North Shore Hospital, Auckland, New to fi rst-line therapy but typically manifests repeated characteristics, and safety profi le, has already been Zealand (D Simpson FRACP); relapses with the need for recurrent therapeutic investigated as maintenance treatment in patients with Helsinki University Central interventions, with disease-free intervals becoming follicular lymphoma.9–13 Previous studies have shown a Hospital, Helsinki, Finland 3,4 (S Leppa MD); progressively shorter. Although some patients can signifi cant clinical benefi t of rituximab maintenance in Institut Bergonié and initially be managed with watchful waiting because they patients with relapsed disease after chemotherapy with Université Victor Segalen are asymptomatic with no adverse prognostic features, or without rituximab.11,12 The use of rituximab Bordeaux 2, Bordeaux, France most need systemic cytotoxic-based treatment. In the maintenance has also been studied after initial (Prof P Soubeyran MD); 5–8 9,10 Academic Medical Centre, past decade, several randomised studies have treatment with single-agent rituximab or chemo- established that the combination of the anti-CD20 therapy alone;13 however, neither of these induction

42 www.thelancet.com Vol 377 January 1, 2011 Articles

regimens is considered optimum as initial treatment suitable than 3 months to reach a trough in serum Amsterdam, Netherlands for patients in need of therapy. Hence, the PRIMA rituximab concentration of 25 μg/mL.14,15 Randomisation (Prof A Hagenbeek MD); CHU de Dijon, Dijon, France (Primary RItuximab and MAintenance) study was was stratifi ed for induction regimen, response to induction (O Casasnovas MD); designed to assess the potential benefi t of 2 years of treatment, geographical region, and centre, with a block Chulalongkorn University, rituximab maintenance after fi rst-line treatment in size of four. Investigators enrolled the participants, and Bangkok, Thailand patients with follicular lymphoma receiving a rituximab assignment to trial groups was done with a computer- (Prof T Intragumtornchai MD); Institut de Cancérologie plus chemotherapy regimen. assisted randomisation allocation sequence (generated by Gustave Roussy, Villejuif, a statistician) that took place centrally at Groupe d’Etude France (C Fermé MD); Methods des Lymphomes de l’Adulte (GELA) central offi ces with a Portuguese Institute of Study design and patients fax process, without the intervention of investigators. Oncology, Lisbon, Portugal (M G da Silva MD); Centre Léon This open-label, international, multicentre randomised Neither the participants nor those giving the interventions, Bérard, Lyon, France study was undertaken between December, 2004, and assessing outcomes, and analysing data were masked to (C Sebban MD); Centre for April 2007, in 223 centres in 25 countries. The trial group assignment. Medical Oncology, Barts and consisted of two phases, induction and maintenance. the London School of Medicine and Dentistry, Queen Mary Patients were eligible for induction if they were older Procedures University of London, London, than 18 years and presented with untreated follicular Initial staging included physical examination; standard UK (Prof A Lister MD); Concord lymphoma (grade 1, 2, or 3a) diagnosed by a lymph-node laboratory assessments; CT scans of the chest, abdomen, Hospital, Concord, NSW, biopsy (done within 4 months of study registration). and pelvis; and bone marrow biopsy. Pathological specimens Australia (J A Estell FRACP); Fundaleu, Buenos Aires, Eligibility required at least one criterion of high tumour were centrally reviewed by a panel of expert pathologists in Argentina (G Milone MD); UCL, burden—namely, bulky disease (one lymphoma lesion each country or in the GELA pathology centre. Mont-Godinne, Yvoir, Belgium greater than 7 cm); three separate nodes of 3 cm or more; During the induction phase, patients were treated with (A Sonet MD); symptomatic splenic enlargement; organ compression by one of three standard immunochemotherapy regimens, F Hoff mann-La Roche, Basel, Switzerland (M Mendila MD); tumour, pleural, or peritoneal eff usion; raised serum with each centre selecting the preferred regimen for all and Centre Henri Becquerel, concentrations of either lactate dehydrogenase or patients enrolled in that centre. The three chemotherapy Rouen, France (Prof H Tilly MD)

β2-microglobulin; or the presence of B symptoms. Patients regimens combined with rituximab were: CVP (cyclo- Correspondence to: had to have a performance status of 2 or less on the Eastern phosphamide 750 mg/m² given intravenously on day 1, Prof Gilles Salles, Hospices Civils Cooperative Oncology Group (ECOG) scale and adequate vincristine 1·4 mg/m² [capped at 2 mg] given de Lyon, Université Claude Bernard, UMR CNRS5239, haematological function (unless due to lymphoma). Non- intravenously on day 1, and prednisone 40 mg/m² given Centre Hospitalier Lyon-Sud, eligibility criteria were a diagnosis of follicular lymphoma orally on days 1–5, with each cycle repeated every 3 weeks 69310 Pierre-Bénite, France grade 3b or transformed into diff use large B-cell lymphoma, for eight cycles), CHOP (cyclophosphamide 750 mg/m² [email protected] CNS involvement, or a life expectancy of less than given intravenously on day 1, vincristine 1·4 mg/m² 6 months. Patients with a previous history of cancer (apart [capped at 2 mg] given intravenously on day 1, doxorubicin For the protocol for the PRIMA from adequately treated non-melanoma skin cancer or in- 50 mg/m² given intravenously on day 1, and prednisone study see http://prima.gela.org situ cervical cancer), with poor renal or hepatic function 100 mg given orally on days 1–5, with each cycle repeated (unless due to lymphoma), or a previous history of allergy every 3 weeks for six cycles), or FCM (fl udarabine to murine products were not eligible. Patients with a 25 mg/m² given intravenously on days 1–3, known HIV infection or an active hepatitis B or hepatitis C cyclophosphamide 200 mg/m² given orally on days 1–3, virus infection were also excluded, but pretreatment testing and mitoxantrone 6 mg/m² given intravenously on day 1, was not initially mandated. We excluded patients if they with each cycle repeated every 4 weeks for six cycles). had had any major surgical procedure or used Rituximab (375 mg/m² for each infusion) was corticosteroids at doses greater than 20 mg per day within administered at day 1 of each chemotherapy course, with 1 month before study entry. two additional infusions administered in patients The protocol was approved by local or national ethics given CHOP (every 3 weeks after the last cycle) and FCM committees according to the laws of each country, and (2 weeks after the fi rst and the fourth cycles) to provide the study was undertaken in accordance with the an equivalent exposure to the antibody during Declaration of Helsinki. Patients were required to provide induction. written informed consent before registration. Response to induction16 was assessed 2–4 weeks after the last induction treatment course. Patients who Randomisation and masking obtained a complete response, an unconfi rmed complete Patients were randomly assigned in a 1:1 ratio to observation response, or a partial response were eligible for or rituximab maintenance (12 infusions of 375 mg/m² randomisation to the maintenance phase of the study. given intravenously, one every 8 weeks) starting 8 weeks Additionally, patients were required to have received at after the last induction treatment. The duration and least six cycles of rituximab plus CVP (R-CVP), four schedule of rituximab maintenance in this study was cycles of rituximab plus CHOP (R-CHOP), or four cycles derived from previous studies using either eight infusions11 of rituximab plus FCM (R-FCM) (each with at least six or 16 infusions13 over 2 years and from pharmacokinetic infusions of rituximab) without a delay of more than studies suggesting that 2-month intervals might be more 2 weeks between each cycle. Any severe underlying www.thelancet.com Vol 377 January 1, 2011 43 Articles

medical disorder that could impair participation or Patients were assessed by clinical examination any major induction treatment-related toxic eff ect every 8 weeks and CT scans every 6 months during precluded eligibility. the 2-year maintenance phase of the study. An end-of-

1217 patients in induction phase

15 excluded (3 sites closed prematurely)

1202 patients registered

885 R-CHOP 272 R-CVP 45 R-FCM

4 withdrawn before treatment 4 withdrawn before treatment 1 withdrawn before treatment

881 patients 268 patients 44 patients treated treated treated

90 withdrawn during treatment 41 withdrawn during treatment 15 withdrawn during treatment 23 major protocol violation 3 major protocol violation 2 major protocol violation 15 treatment failure 12 treatment failure 4 treatment failure 17 treatment toxicity 5 treatment toxicity 5 treatment toxicity 1 patient voluntary 1 patient voluntary 1 patient voluntary withdrawal withdrawal withdrawal 5 death 1 death 3 other 29 other 19 other

791 completed 227 completed 29 completed treatment treatment treatment

22 not randomised 5 not randomised 1 not randomised 1 major protocol violation 1 patient voluntary 1 treatment toxicity 3 treatment toxicity withdrawal 3 patient voluntary 4 other withdrawal 15 other Figure 1: Trial profi le When the reasons for 769 patients 222 patients 28 patients withdrawal categorised as randomised randomised randomised other were investigated and centrally reviewed, of the 184 patients withdrawing 1019 randomised to from registration until maintenance randomisation, the main phase reasons leading to withdrawal were: major protocol violation 1 excluded (including eligibility criteria or Patient died during inadequate induction randomisation treatment, n=61); complete or partial response not achieved 1018 ITT population after induction therapy (n=46); toxic eff ects or delays 513 observation 505 rituximab during induction treatment administration (n=44); 5 withdrawn before treatment 4 withdrawn before treatment underlying medical disease or death (n=18); patient decision 508 patients 501 patients (n=8); and investigator observed treated decision (n=7). R-CHOP=rituximab, 166 withdrawn during observation 103 withdrawn during observation 147 treatment failure 69 treatment failure cyclophosphamide, 1 treatment toxicity 10 treatment toxicity doxorubicin, vincristine, and 8 patient voluntary 4 not assessed at 5 not assessed at 10 patient voluntary withdrawal prednisone. R-CVP=rituximab, withdrawal end of end of 1 death cyclophosphamide, vincristine, 10 other observation observation 13 other and prednisone. R-FCM=rituximab, fl udarabine, 338 patients 393 patients cyclophosphamide, and completed completed mitoxantrone. ITT=intention observation treatment to treat.

44 www.thelancet.com Vol 377 January 1, 2011 Articles

treatment assessment also required a bone marrow Patients who Randomised patients biopsy if initially involved. Follow-up assessments received induction included clinical evaluation every 3 months and CT treatment scans every 6 months for an additional 3 years. (n=1193) Quality-of-life questionnaires for functional Observation Rituximab assessment of cancer therapy—general (FACT-G) (n=513) maintenance score and European Organisation for Research and (n=505) Treatment of Cancer (EORTC) QLQ-C30 scale were Age >60 years 423 (35%) 180 (35%) 176 (35%) planned to be completed by patients at registration, at Age (years) 56 (22–87) 55 (22–84) 57 (26–79) the end of induction, and then every 12 months. Male sex 622 (52%) 263 (51%) 270 (53%) Ann Arbor stage III/IV 1075 (90%) 459 (89%) 459 (91%) Statistical analysis ECOG performance status ≥1 434 (36%) 172 (34%) 181 (36%) The study was designed to show a 45% increase in B symptoms present 388 (33%) 156 (30%) 160 (32%) median progression-free survival (PFS) from the time of Bone marrow lymphoma involvement 654 (55%) 285 (56%) 275 (54%) randomisation (6 months after the start of induction Lactate dehydrogenase >ULN* 403 (34%) 164 (32%) 173 (34%) therapy) with a power of 80% and an overall two-sided Haemoglobin <120 g/L 239 (20%) 96 (19%) 100 (20%) type I error of 5%, with use of a two-sided log-rank test. β2 microglobulin ≥3 mg/L* 348 (32%) 132 (28%) 148 (32%) The trial was intended in 2004 to register 640 patients FLIPI score† and randomly assign 480 (assuming that 75% of the Low (0–1 risk factors) 254 (21%) 110 (21%) 106 (21%) patients would be randomly assigned on the basis of Intermediate (2 risk factors) 423 (36%) 187 (36%) 183 (36%) response rate and eligibility criteria). Mature information High (3–5 risk factors) 514 (43%) 216 (42%) 215 (43%) 11,17,18 then became available from earlier trials, indicating a Initial local diagnosis of FL (other than grade 3B) 1188 (100%) 512 (100%) 504 (100%) 6-month delay before seeing a benefi t of maintenance Central pathological review done 1115 (93%) 487 (95%) 467 (92%) after immunochemotherapy, and patient accrual was Confi rmed FL (other than grade 3B) 994 (84%) 433 (84%) 425 (84%) more rapid than was expected. Therefore, two protocol Diagnosis of other lymphoma subtype‡ 56 (5%) 28 (5%) 16 (3%) amendments were implemented before the fi rst data Unclassifi able or not assessable for technical reasons 65 (6%) 26 (5%) 26 (5%) analysis, increasing the fi nal sample size to 1200 patients Induction immunochemotherapy regimen registered and 900 randomly assigned, and allowing a R-CHOP 885 (74%) 386 (75%) 382 (76%) more meaningful examination of the primary endpoint R-CVP 272 (23 %) 113 (22%) 109 (22%) in subgroups. Two interim analyses were originally R-FCM 45 (4%) 14 (3%) 14 (3%) scheduled to be done after 50% (n=172) and 75% (n=258) Response to induction of the total number of anticipated events (n=344). The Complete response .. 195 (38%) 205 (41%) protocol was subsequently amended to remove the fi rst Unconfi rmed complete response .. 165 (32%) 155 (31%) interim analysis because the results would have been Partial response .. 152 (30%) 139 (28%) regarded as clinically immature with short follow-up. Other§ .. 1 (≤1%) 6 (1%) The α-spending function with the O’Brien–Fleming boundary was applied for the interim analysis to maintain Data are number (%) or median (range). ECOG=Eastern Cooperative Oncology Group. ULN=upper limit of normal. the overall two-sided type I error of 0·05. Subgroup FLIPI=follicular lymphoma international prognostic index. FL=follicular lymphoma. R-CHOP=rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone. R-CVP=rituximab, cyclophosphamide, vincristine, and analyses for PFS were planned according to age, sex, prednisone. R-FCM=rituximab, fl udarabine, cyclophosphamide, and mitoxantrone. *Lactate dehydrogenase and categories defi ned by the initial follicular lymphoma β2-microglobulin serum concentrations available for only 1188 and 1101 patients, respectively. †FLIPI scores were international prognostic index (FLIPI),19 induction collected at registration. The risk score includes fi ve factors: age (>60 years), Ann Arbor stage (III or IV), haemoglobin 17 regimen, and response to induction (complete or (≤120 g/L), serum lactate dehydrogenase (>ULN), and number of nodal areas involved (fi ve or more). Data are available for 1191 patients. ‡Other lymphoma subtypes enrolled included FL grade 3B (n=12), combined diff use large unconfi rmed complete response, or partial response). B-cell lymphoma and FL (n=24), diff use large B-cell lymphoma (n=10), mantle cell lymphoma (n=6), small lymphocytic The primary study endpoint was PFS from the time of lymphoma (n=2), Hodgkin’s lymphoma (n=1), and angioimmunoblastic T-cell lymphoma (n=1). §After data cleaning, randomisation to rituximab maintenance or no further one patient in the observation group and four in the rituximab maintenance group were assessed as having stable disease at randomisation, and two other patients in the rituximab maintenance group missed standard assessment treatment (observation). Secondary endpoints were procedures. event-free survival, time to next chemotherapy treatment, time to next antilymphoma treatment, overall survival, Table 1: Baseline demographics and clinical characteristics of patients at enrolment (before immunochemotherapy induction treatment), and at randomisation, according to study group response rate at the end of maintenance, safety, toxic eff ects, and quality of life. Response and progression were defi ned with and response to induction) in the assessment of international standard criteria.16 Survival functions were maintenance eff ect. Response rates and frequency of estimated by the Kaplan-Meier method and compared by adverse events were compared with the χ² test. To log-rank test stratifi ed by induction regimen and establish whether the study treatment groups diff ered in induction treatment response. Cox regression analysis quality of life assessed at the end of treatment, was done to adjust for the eff ect of known prognostic independent of any diff erences in quality of life assessed factors (age, sex, FLIPI category, induction treatment, at the end of induction, we used ANCOVA. FACT-G total www.thelancet.com Vol 377 January 1, 2011 45 Articles

scores and EORTC QLQ-C30 global health scores were groups, F Hoff mann-La Roche, and Biogen IDEC. All collected during rituximab maintenance or observation, case report forms were sent to the GELA central with repeated measurements over time (censored at the operation offi ce (GELA-RC), and double data entry was date of disease progression to avoid potential biases undertaken for verifi cation purposes. Queries and on- related to toxic eff ects of second-line treatment), and site monitoring data were used for fi nal validation, and analysed with an unstructured covariance model. All GELA-RC had full control of the database. An effi cacy analyses were done in the intention-to-treat independent data safety monitoring committee randomised population. examined the safety data every year, advised on protocol On Jan 14, 2009, 267 events were recorded, leading to amendments, and undertook the prescheduled interim the protocol-specifi ed interim analysis undertaken by analysis of the effi cacy results. Final statistical analyses the data safety monitoring committee that declared the were done independently by GELA-RC and F Hoff mann- primary endpoint of PFS had been met and that the La Roche. The corresponding author was responsible study should be fully analysed. The study was then for data analysis, data interpretation, and writing of the terminated; at that time the median follow-up after report; had full access to all the data in the study; and randomisation was 25 months (IQR 19–30) with had the fi nal responsibility for the decision to submit 231 patients still to complete the maintenance or for publication. observation period. This report describes the analysis done after an additional year of follow-up, with a cutoff Results date of Jan 15, 2010, when all randomised patients had Figure 1 shows the trial profi le. On-site monitoring reached the time for end-of-treatment evaluation or showed that three centres did not adhere to good clinical were withdrawn. practice and were closed (excluding 15 patients); nine The study is registered with ClinicalTrials.gov, other registered patients withdrew before receiving any number NCT00140582. treatment. Therefore, 1193 patients had complete data at study enrolment (table 1). Of those, 1115 (93%) had Role of the funding source central pathology review done, and follicular lymphoma The GELA acted as the sponsor of the study and (except grade 3B) was confi rmed in 994 cases (89%). We developed and undertook the study protocol in noted no diff erences in demographics or disease collaboration with several other lymphoma study characteristics at enrolment between patients who

A B 1·0 Rituximab maintenance Observation 0·8

0·6

0·4 Event-free rate Event-free 0·2 HR 0·55 (95% CI 0·44–0·68); p<0·0001 HR 0·60 (95% CI 0·47–0·76); p<0·0001 0 Time (months) Time (months) Number at risk Rituximab 505 472 445 423 404 307 207 84 17 0 505 483 455 441 414 312 209 91 17 0 Observation 513 469 415 367 334 247 161 70 16 0 513 487 452 417 380 286 170 71 18 0

C D 1·0

0·8

0·6

0·4 Event-free rate Event-free 0·2 HR 0·62 (95% CI 0·47–0·81); p=0·0004 HR 0·87 (95% CI 0·51–1·47); p=0·60 0 0 6 12 18 24 30 36 42 48 54 60 0 6 12 18 24 30 36 42 48 54 60 Time (months) Time (months) Number at risk Rituximab 505 484 459 444 428 325 220 93 19 0 505 499 492 483 474 365 246 108 22 1 Observation 513 492 460 425 393 302 188 75 20 0 513 507 501 492 472 381 243 97 26 0

Figure 2: Kaplan-Meier estimates of progression-free survival (A), time to next antilymphoma treatment (B), time to next chemotherapy (C), and overall survival (D) from randomisation with rituximab maintenance versus observation HR=hazard ratio.

46 www.thelancet.com Vol 377 January 1, 2011 Articles

received R-CHOP (n=885), R-CVP (n=272), or unconfi rmed complete response compared with 268 of R-FCM (n=45) induction (data not shown). 146 patients 513 patients (52·2%, 47·8–56·6) in the observation group withdrew during induction treatment, a further (estimated diff erence 18·0%, 12·3–23·6; p=0·0001). More 28 completed induction but were not randomly assigned patients who were in partial response at the time of (fi gure 1), and one patient died before notifi cation of the randomisation converted to complete or unconfi rmed randomisation. Therefore, from randomisation, the complete response after 2 years in the rituximab intention-to-treat population consisted of 1018 patients maintenance group (72/139 [52%]) than did those in (505 in the rituximab maintenance group and 513 in the the observation group (45/152 [30%]; estimated observation group; table 1). diff erence 22·2%, 95% CI 11·2–33·3, p=0·0001). With a median follow-up of 36 months in both groups (IQR 30–42), 130 of 505 patients in the rituximab HR (95% CI) N maintenance group and 218 of 513 in the observation group had documented disease progression, and fi ve and All 0·55 (0·44–0·68) 1018 Age (years) three patients, respectively, had died before disease <60 0·49 (0·37–0·65) 624 progression. 3-year PFS was 74·9% (95% CI 70·9–78·9) ≥60 0·67 (0·47–0·94) 394 in the rituximab maintenance group and 57·6% Sex (53·2–62·0) in the observation group (stratifi ed log rank, Women 0·63 (0·45–0·87) 485 p<0·0001; fi gure 2A). The risk of progression was Men 0·48 (0·36–0·64) 533 signifi cantly reduced for the rituximab maintenance FLIPI index group (hazard ratio [HR] 0·55, 95% CI 0·44–0·68). The ≤1 0·39 (0·21–0·72) 216 median time to progression was not reached in the 2 0·44 (0·30–0·64) 370 rituximab maintenance group and was estimated to be ≥3 0·68 (0·51–0·92) 431 48·3 months (95% CI 38·0 to not reached) in the Induction chemotherapy R-CHOP 0·51 (0·39–0·65) 768 observation group. R-CVP 0·68 (0·45–1·02) 222 Pre-planned analyses of patient subgroups categorised R-FCM 0·54 (0·13–2·24) 28 by age, sex, FLIPI score category, induction chemotherapy, Response to induction and response to induction showed that the eff ect of CR/CRu 0·57 (0·44–0·74) 720 rituximab maintenance was consistent across these PR 0·48 (0·32–0·72) 291 diff erent subgroups, although with borderline results for 0 1 2 3 patients having received R-CVP (number for those having Favours maintenance Favours observation received R-FCM are too small to conclude) (fi gure 3). In a Cox regression multivariate analysis adjusted by prognostic Figure 3: Risk of progression with rituximab maintenance versus observation, according to prespecifi ed subgroups factors, a longer PFS was signifi cantly associated with HR=hazard ratio. FLIPI=follicular lymphoma international prognostic index. R-CHOP=rituximab, randomisation to the rituximab maintenance group cyclophosphamide, doxorubicin, vincristine, and prednisone. R-CVP=rituximab, cyclophosphamide, vincristine, (HR 0·55, 95% CI 0·44–0·68; p<0·0001), an age of and prednisone. R-FCM=rituximab, fl udarabine, cyclophosphamide, and mitoxantrone. CR=complete response. 60 years or older (0·68, 0·54–0·86; p=0·0013), female CRu=unconfi rmed complete response. PR=partial response. sex (0·76, 0·62–0·94; p=0·013), lower FLIPI score categories (overall p<0·0001), and R-CHOP or R-FCM as Observation (n=508) Rituximab maintenance (n=501) induction treatment (0·39, 0·17–0·89; p=0·0029). Grade 3/4 Leading to treatment Grade 3/4 Leading to treatment Overall, 102 of 505 patients in the rituximab maintenance discontinuation discontinuation group and 167 of 513 patients in the observation group All adverse events 84 (17%) 8 (2%) 121 (24%) 19 (4%)† had begun a new antilymphoma treatment, which was a Neoplasia 17 (3%) 6 (1%) 20 (4%) 5 (1%) chemotherapy regimen in 80 patients in the rituximab Neutropenia 5 (1%) 0 18 (4%) 0 maintenance group and 129 in the observation group. We Febrile neutropenia 2 (<1%) 0 1 (<1%) 1 (<1%) recorded signifi cant reductions in the risk of starting a Infections 5 (1%) 0 22 (4%) 4 (1%) new antilymphoma treatment or death (fi gure 2B) or CNS disorders 13 (3%) 0 10 (2%) 0 starting a new chemotherapy or death (fi gure 2C) in the Cardiac disorders 5 (1%) 0 11 (2%) 1 (<1%) rituximab maintenance group. Event-free survival was Pregnancy NA 2 (<1%) NA 3 (1%) also signifi cantly improved in the rituximab maintenance Data are number (%). NA=not applicable. *Safety during maintenance was assessed for patients who undertook at group (stratifi ed HR 0·59, 95% CI 0·48–0·72). With only least one visit (rituximab treatment or observation) after randomisation. All adverse events, defi ned as any adverse 26 deaths recorded in the rituximab maintenance group change from the patient’s baseline condition, whether considered related to treatment or not, were collected and and 30 in the observation group with present follow-up, graded according to the Common Terminology Criteria for Adverse Events 3.0 grading system.18 All grade 3 and we noted no signifi cant diff erence in overall survival 4 events plus grade 2 infections were recorded in detail during maintenance or observation and 6 months thereafter. †Other events leading to treatment discontinuation were pyrexia, fulminant hepatitis, hypersensitivity, (fi gure 2D). post-procedural fi stula, and lung disorder (one case each). At the end of the maintenance phase of the study, 361 of 505 patients (71·5%; 95% CI 67·3–75·4) in the Table 2: Grade 3 and 4 adverse events* experienced by 2% or more of patients and adverse events leading to treatment discontinuation, after randomisation to rituximab maintenance or observation rituximab maintenance group were in complete or www.thelancet.com Vol 377 January 1, 2011 47 Articles

Of the 1009 patients assessed for safety, adverse events Baseline (randomisation) After 1 year End of maintenance phase were reported in 281 of 501 (56%) patients in the rituximab Maintenance Observation Maintenance Observation Maintenance Observation maintenance group and 189 of 508 (37%) in the observation IgA group (risk ratio 1·51, 95% CI 1·32–1·73; p<0·0001). The n 131 125 118 89 111 61 most common adverse events reported were grade g/L 1·33 (0·63) 1·57 (3·55) 1·26 (0·58) 1·40 (0·88) 1·25 (0·50) 1·55 (0·91) 2–4 infections in 197 of 501 (39%) patients and 123 of IgG 508 (24%) patients, respectively (risk ratio 1·62, 1·35–1·96; n 131 125 118 89 100 60 p<0·0001). The fi ve most common infections reported in g/L 7·87 (2·26) 7·76 (2·13) 7·73 (2·12) 8·22 (2·22) 7·48 (2·13) 8·31 (2·32) the rituximab and observation groups were bronchitis IgM (54 and 28 cases, respectively), upper respiratory tract n 127 121 113 89 110 61 infections (28 and 11 cases), sinusitis (21 and eight cases), g/L 0·64 (0·52) 0·59 (0·55) 0·55 (0·39) 0·55 (0·31) 0·51 (0·39) 0·58 (0·25) urinary tract infections (14 and nine cases), nasopharyngitis (11 and 14 cases), and urinary tract infections (14 and Immunoglobulin concentrations are given as mean (SD). Serum concentrations of immunoglobulins were assessed nine cases), whereas the cumulative number of herpes only in a subset of patients in one participating country. viruses-related infections were 19 and 12, respectively. Table 3: Serum concentrations of immunoglobulins at randomisation, after 1 year of maintenance, and Grade 3 or 4 adverse events20 (table 2) occurred in 121 of at the end of the maintenance phase 501 (24%) patients in the rituximab group and 84 of 508 (17%) patients in the observation group (risk ratio 1·46, 1·14–1·87; p=0·0026). 19 (4%) and eight (2%) events, respectively, resulted in treatment discontinuation (risk A ratio 2·41, 1·06–5·45; p=0·029). 120 Rituximab maintenance Observation Only one death (fulminant hepatitis B in the absence of viral suppressive therapy in the rituximab maintenance 100 group) was reported to be possibly related to treatment toxic eff ects; other causes of deaths before lymphoma 80 recurrence were attributed to other malignant diseases (four cases; one in rituximab maintenance group and three 60 in observation group), or pulmonary haemorrhage,

40 accident, or unknown (sudden death) (one case each in rituximab maintenance group). Two other patients (one Mean global health status score 20 from each group) developed progressive multifocal Induction Maintenance Follow-up leukoencephalopathy after lymphoma relapse and

0 subsequent treatments, which included investigational Number of patients agents for both patients. At the end of 2 years of rituximab Rituximab maintenance 394 343 217 224 93 maintenance or observation, median serum concentrations Observation 374 341 166 158 56 of immunoglobulin isotypes did not diff er signifi cantly B between the rituximab maintenance and observation 120 groups (table 3). The mean adjusted FACT-G total scores at the end of 100 treatment were 86·6 (95% CI 85·0–88·3) in the rituximab maintenance group and 87·2 (85·3–89·1) in 80 the observation group, suggesting no association of these scores with treatment group (ANCOVA adjusted 60 for scores at the end of induction, p=0·68). Consistently, the EORTC QLQ-C30 global health status

Mean total score 40 mean scores were 75·5 (72·8–78·2) and 75·2 (72·0–78·4), respectively (p=0·89; fi gure 4). The analysis 20 Induction Maintenance Follow-up of repeated measures was consistent with ANCOVA, and did not show any statistical diff erence on either 0 scale (data not shown). Baseline After induction 1 year 2 years 3 years Number of patients Rituximab maintenance 384 334 213 219 96 Discussion Observation 375 331 168 160 54 Results of the PRIMA study show that 2 years of rituximab maintenance therapy signifi cantly prolongs Figure 4: Quality of life during 2 years of treatment with rituximab maintenance versus observation alone PFS, delays the time to next antilymphoma treatment (A) European Organisation for Research and Treatment of Cancer quality-of-life questionnaire C30 scale. (B) Functional assessment of cancer therapy—general scale. Quality-of-life assessments were censored at time of and next chemotherapy, and improves the quality of progression. response in patients with previously untreated follicular

48 www.thelancet.com Vol 377 January 1, 2011 Articles

lymphoma that is responsive to fi rst-line rituximab plus chemotherapy. Rituximab maintenance was well Panel: Research in context tolerated, with a limited number of adverse events Systematic review resulting in treatment discontinuation, and there were We searched Medline from January, 1995, to no unexpected safety fi ndings. In line with other November, 2010, for full papers reporting randomised clinical studies,21 we recorded a signifi cantly increased incidence trials and meta-analyses with the terms “rituximab of infectious events, mostly of mild to moderate severity, maintenance” and “lymphoma”. We identifi ed fi ve despite no signifi cant decrease in serum randomised clinical trials in patients with follicular immunoglobulin concentrations. Physicians and lymphoma,9–13 with two updates22,33 and one meta-analysis.21 patients should be aware of this risk to optimise the management of these patients. Further follow-up of Interpretation immunoglobulin concentrations will also be done in Together, these studies provide substantial evidence that the trial. Nevertheless, despite the higher frequency of rituximab maintenance improves the outcome of patients adverse events in the rituximab maintenance group with follicular lymphoma in term of progression-free and than in the observation group, few patients withdrew overall survival. The use of rituximab maintenance is from the study because of toxic eff ects. Furthermore, associated with an increased risk of infections. Our results are the burden associated with repeated infusions over consistent with fi ndings from other studies, but provide 2 years did not seem to impair patient quality of life, evidence that this intervention improves progression-free which was similar in both study groups. However, since survival and response rate in patients with follicular only a subset of patients completed the quality-of-life lymphoma responding to a combination of chemotherapy questionnaires, these data should be interpreted plus rituximab administered as fi rst-line treatment. cautiously, because we cannot exclude a reporting bias favouring patients who did not have treatment-related healthy individuals26 and in patients,27,28 where they can adverse events. lead to disease recurrence. Rituximab maintenance The signifi cant reduction in rate of lymphoma might eventually exert a long-term control on progression in the rituximab maintenance group these cells. was consistent between patients with diff erent Follicular lymphoma is an indolent disease, with demographics, disease characteristics, and prognostic prolonged survival even in cases that are ultimately factors, within the limitations of the study eligibility fatal. The effi cacy of salvage therapies used after initial criteria. Together with the results of the multivariate Cox treatment failure could also preclude the demonstration regression analysis, these data suggest that all analysed of an overall survival benefi t associated with fi rst-line categories of patients eligible for fi rst-line immuno- therapy.22,29–31 With the present 3-year follow-up, less chemotherapy benefi ted from rituximab maintenance. than 5% of patients in either group had died, with no Notably, we recorded a reduction in the risk of lymphoma signifi cant diff erence between the two groups, and recurrence irrespective of the intensity of the fi rst-line salvage therapy results are still immature. However, a induction therapy and the response achieved. Although higher proportion of patients achieved a complete we noted no gain in overall survival, the reduction in the response at the end of rituximab maintenance risk of disease progression after responding to induction treatment than in the observation group. Attainment of is likely to be preferred by patients with follicular a complete response is associated with improved long- lymphoma. This preference should be balanced with the term survival in patients with follicular lymphoma.32 constraints and costs associated with 2-year rituximab Since longer follow-up will be needed to show any maintenance. However, in view of the substantial possible eff ect of rituximab maintenance on overall improvement in patient survival during the past decade2 survival, we will continue to follow up these patients. and the indication that rituximab maintenance might also The results of other studies assessing radio- result in overall survival benefi ts in relapsed follicular immunotherapy con solidation with iodine-131 labelled lymphoma,11,21,22 we should not exclude such a possible tositumomab (regis-tered with ClinicalTrials.gov, benefi t of this intervention in the long term. numbers NCT00006721 and NCT00770224) or rituximab Diff erent mechanisms of action have been proposed maintenance for 4 or 5 years for the therapeutic activity of rituximab.23 When used as (NCT00877214 and NCT00227695) might also provide a single agent in a prolonged maintenance scheme, the some insights about how to improve the outcome of immune-mediated activity of the antibody could patients with follicular lymphoma. possibly be more potent than other mechanisms. Data In summary, the data from this study suggest that suggest that anti-CD20 might induce a rituximab maintenance in patients with high tumour T-cell-specifi c response against lymphoma cells.24,25 burden follicular lymphoma, who respond to rituximab Other studies investigating the mechanisms of follicular plus chemotherapy induction, improves PFS and should lymphoma development have suggested that lymphoma now be considered as fi rst-line treatment for these precursor cells are able to survive for years both in patients (panel). www.thelancet.com Vol 377 January 1, 2011 49 Articles

Contributors paid employee of Roche and has stock options with Roche. GS is also GS, JFS, FO, AL-G, LX, PF, PB, CH, AH, AL, AS, MM, BC, and HT an advisory board member for Celgene, Janssen-Cilag, Genzyme, and designed the study. GS, JFS, FO, AL-G, DB, LX, LMP, SL, TI, MGdS, Calistoga; and has received travel support from Pfi zer and support for AH, AL, and MM were responsible for the conduct of the study centrally educational materials from Janssen-Cilag and Celgene. ALG has or at the country-specifi c level. GS, JFS, FO, AL-G, LX, PF, PB, CH, received support for educational materials from Celgene. HT is an LMP, SL, TI, MGdS, RB, JVC, DC, AD, DS, PS, OC, CF, CS, JAE, GM, advisory board member for Celgene and Seattle Genetics; has received AH, AL, AS, MM, BC, and HT contributed research data to the study. grant support from Celgene and Amgen; and has received support for GS, JFS, FO, AL-G, LX, PF, PB, CH, AH, AL, AS, BC, and HT preparation of educational materials from Celgene, Amgen, contributed to data analysis and interpretation. GS drafted the report, Janssen-Cilag, and Pfi zer. SL declared consultancy and honoraria for which all co-authors critically revised for signifi cant scientifi c content. Bayer and has received travel support from GlaxoSmithKline. FO, LX, JVC, RB, PB, DC, LMP, DS, TI, CF, CS, JE, GM, and AS declare that Investigators participating in the study they have no confl icts of interest. France G Salles, P Feugier, R Bouabdallah, C Gisselbrecht, C Haioun, A Delmer, H Tilly, O Casasnovas, C Ferme, P Soubeyran, C Sebban, Acknowledgments B Christian, R Delarue, D Guyotat, H Maisonneuve, O Fitoussi, This study was funded by Groupe d’Etude des Lymphomes de l’Adulte J Gabarre, T Lamy, F Morschhauser, J F Rossi, D Decaudin, (GELA; Paris, France) and F Hoff mann-La Roche (Basel, Switzerland). P Colombat, V Delwail, M Janvier, C Recher, C Salanoubat, We thank the pathologists undertaking central case review, particularly Z Marjanovic, M Blanc, C Foussard, J-L Harrousseau, E Jourdan, N Brousse, D Canioni, F Charlotte, C Chassagne-Clément, P Dartigues, F Maloisel, L Al Jassem, T De Revel, A Devidas, J C Eisenmann, B Fabiani, L Deleval, E Campos, and D DeJong; the statisticians E Fleck, G Lepeu, C Martin, B Corront, P Moreau, A Thyss, B Anglaret, C Bergé, M Fournier, J Maurer, and their teams; the entire GELA-RC B Salles, M Alexis, K Bouabdallah, S Castaigne, F Dreyfus, P Fenaux, team, including D Germain as project manager, all monitors and C Fruchart, M Macro, F Bauduer, D Bordessoule, M Fabbro, clinical research associates, and the data management and A Le Pourhiet, S Sadot Le Bouvier, P Solal Celigny, X Vallantin, pharmacovigilance teams; the teams from the Australasian Leukaemia C Kulekci, S Lefort, L Mosser, J F Ramee, N Morineau, B Audhuy, and Lymphoma Group, the Fundación Farreras Valentí, the Czech F Boue, M Flesch, H Gonzalez, J Gutnecht, F Marechal, A Belhabri, Lymphoma Study Group, the Hemato-Oncologie Volwassenen W Abarah, S Cailleres, N Denizon, O Fain, J-M Karsenti, P Morel, Nederland (HOVON), and the UK Haematology Trials Group for their J-N Munck, H Cure, O Tournilhac, M Wetterwald. Australia J Catalano, contribution in organising the study in their respective countries; the J Estell, N Wickham, P Marlton, J Seymour, M Walsh, P Bardy, Roche teams for their support and contribution to the study; the Data U Hahn, M Hertzberg, D Ma, I Prosser, C Tiley, R Filshie, C Arthur, Safety Monitoring Committee members J O Armitage, D Hasenclever, K Fay, P Campbell, G Kannourakis, J Bashford, R Blum, R Herrmann, and M Ghielmini; and R Marcus for his original input into the study I Irving, M Leahy, I Lewis, R Lowenthal, J McKendrick, A Spencer, design. Editing assistance for the manuscript was provided by C Underhill, T Brighton, G Cull, B Augustson. Y-L Kwan. John Carron, an employee of an independent medical writing agency Belgium F Off ner, A Bosly, P Zachee, M Maerevoet, T Connerotte, funded by F Hoff mann-La Roche. E Van Den Neste, A Van Hoof, K Van Eygen, B De Prijck, References S Van Steenweghen, D Bron, A Kentos, P Pierre, H Demuynck, 1 Swenson WT, Wooldridge JE, Lynch CF, Forman-Hoff man VL, M Andre, O Hamdam, V Mathieux, P Mineur, V Verschaeve. Chrischilles E, Link BK. Improved survival of follicular lymphoma Spain D Caballero, E Montserrat, A López Guillermo, J J Bargay, patients in the United States. J Clin Oncol 2005; 23: 5019–26. S Gardella, J A Marquez, J A Soler, J Briones, C Estany, J Besalduch, 2 Pulte D, Gondos A, Brenner H. Ongoing improvement in outcomes J Capote, L Escoda, E Monzo, E Perez Ceballos, J Ribera, A Bailen, for patients diagnosed as having Non-Hodgkin lymphoma from I Espagnol, F Losal, J A Garcia Marco, A Lopez, J M Macia Virgili. the 1990s to the early 21st century. Arch Intern Med 2008; Denmark L Moller Pedersen, P B Hansen, O Vestergaard Gadeberg, 168: 469–76. M Hansen, I B Sillesen, A Bukh, H Hasselbalch, N Toff ner-Clausen, 3 Gallagher CJ, Gregory WM, Jones AE, et al. Follicular lymphoma: L Enggaard, F D’Amore, S Ingeberg, M Petersen, A J Vangsted. prognostic factors for response and survival. J Clin Oncol 1986; Czech Republic A Oborilova, D Belada, M Trneny, T Papajik, 4: 1470–80. H Siff nerova, M Brejcha, M Jankovska, M Matuska, J Prausova. 4 Johnson PW, Rohatiner AZ, Whelan JS, et al. Patterns of survival in New Zealand D Simpson, L Berkahn, P Ganly, H Blacklock. patients with recurrent follicular lymphoma: a 20-year study from Finland S Leppa, K Vasala, T Lehtinen, S Jyrkkio. a single center. J Clin Oncol 1995; 13: 140–47. Netherlands A Hagenbeek, J W Baars, M Chamuleau, M R Schipperus, 5 Marcus R, Imrie K, Belch A, et al. CVP chemotherapy plus D H Biesma, O De Weerdt, R E Brouwer, F Heyning, rituximab compared with CVP as fi rst-line treatment for advanced J M M Raemaekers, P Sonneveld, J K Doorduijn, R Van der Griend, follicular lymphoma. Blood 2005; 105: 1417–23. M B Van’t Veer, E Lugtenburg. Thailand I Tanin, C Suporn, 6 Hiddemann W, Kneba M, Dreyling M, et al. Frontline therapy with S Noppadol. Portugal M Gomes da Silva. UK R Pettengell, A Lister, rituximab added to the combination of cyclophosphamide, K Ardeshna, D Culligan, D Cunningham, P Johnson, H McCarthy, doxorubicin, vincristine, and prednisone (CHOP) signifi cantly improves the outcome for patients with advanced-stage follicular D Moir, D White, J Radford. lymphoma compared with therapy with CHOP alone: results of Argentina G Milone, J Milone, G Kusminsky, D Riveros. a prospective randomized study of the German Low-Grade Brazil A Zanichelli, C Chiattone, D Chamone, J Vaz, D Tabak. Lymphoma Study Group. Blood 2005; 106: 3725–32. Colombia M Urrego, A Zambrano, M Duarte. India G Babu, A Ranade, 7 Herold M, Haas A, Srock S, et al. Rituximab added to fi rst-line A Vaid, D C Doval. Peru L Casanova, F Hurtado de Mendoza. mitoxantrone, chlorambucil, and prednisolone chemotherapy Israel O Shpilberg, D Attias, I Levi. Serbia D Boskovic, S Popovic, followed by interferon maintenance prolongs survival in patients D Jovanovic. Venezuela M A Torres. China Z Yongqiang, S Zhi-Xiang, with advanced follicular lymphoma: an East German Study G Zhong-Zhen, S Yuan Kai. Croatia J Branimir. Turkey B Ferhanoglu, Group Hematology and Oncology Study. J Clin Oncol 2007; M Cetin, M Ozcan, B Undar. Uruguay E Bodega, M Nesse. 25: 1986–92. 8 Salles G, Mounier N, de Guibert S, et al. Rituximab combined with Confl icts of interest chemotherapy and interferon in follicular lymphoma patients: results GS, JFS, ALG, DB, PF, PS, CH, SL, AH, OC, AL, BC, AD, and HT of the GELA-GOELAMS FL2000 study. Blood 2008; 112: 4824–31. declared consultancy, honoraria, and/or advisory board membership 9 Hainsworth JD, Litchy S, Burris HA 3rd, et al. Rituximab as from Roche and/or Genentech. PF, CH, AD, SL, PS, OC, AL, and BC fi rst-line and maintenance therapy for patients with indolent received grant support from Roche. GS, JFS, DB, CH, AD, SL, PS, non-Hodgkin’s lymphoma. J Clin Oncol 2002; 20: 4261–67. MGdS, and BC received travel support from Roche. GS, JFS, ALG, SL, 10 Ghielmini M, Schmitz SF, Cogliatti SB, et al. Prolonged treatment and AL have received support for preparation of educational materials with rituximab in patients with follicular lymphoma signifi cantly and/or lectures from Roche. PF has received payment for participation increases event-free survival and response duration compared with in data review meetings and has stock options with Roche. MM is a the standard weekly x 4 schedule. Blood 2004; 103: 4416–23.

50 www.thelancet.com Vol 377 January 1, 2011 Articles

11 van Oers MH, Klasa R, Marcus RE, et al. Rituximab maintenance 22 van Oers MH, Van Glabbeke M, Giurgea L, et al. Rituximab improves clinical outcome of relapsed/resistant follicular maintenance treatment of relapsed/resistant follicular non-Hodgkin’s non-Hodgkin lymphoma in patients both with and without lymphoma: long-term outcome of the EORTC 20981 phase III rituximab during induction: results of a prospective randomized randomized intergroup study. J Clin Oncol 2010; 28: 2853–58. phase 3 intergroup trial. Blood 2006; 108: 3295–301. 23 Cartron G, Watier H, Golay J, Solal-Céligny P. From the bench to 12 Forstpointner R, Unterhalt M, Dreyling M, et al. Maintenance the bedside: ways to improve rituximab effi cacy. Blood 2004; therapy with rituximab leads to a signifi cant prolongation of 104: 2635–42. response duration after salvage therapy with a combination of 24 Hilchey SP, Hyrien O, Mosmann TR, et al. Rituximab rituximab, fl udarabine, cyclophosphamide, and mitoxantrone immunotherapy results in the induction of a lymphoma (R-FCM) in patients with recurring and refractory follicular and idiotype-specifi c T-cell response in patients with follicular mantle cell : Results of a prospective randomized study lymphoma: support for a «vaccinal eff ect» of rituximab. Blood 2009; of the German Low Grade Lymphoma Study Group (GLSG). 113: 3809–12. Blood 2006; 108: 4003–08. 25 Abès R, Gélizé E, Fridman WH, Teillaud JL. Long-lasting antitumor 13 Hochster H, Weller E, Gascoyne RD, et al. Maintenance rituximab protection by anti-CD20 antibody through cellular immune after cyclophosphamide, vincristine, and prednisone prolongs response. Blood 2010; 116: 926–34. progression-free survival in advanced indolent lymphoma: results 26 Roulland S, Navarro JM, Grenot P, et al. Follicular lymphoma-like of the randomized phase III ECOG1496 Study. J Clin Oncol 2009; B cells in healthy individuals: a novel intermediate step in early 27: 1607–14. lymphomagenesis. J Exp Med 2006; 203: 2425–31. 14 Gordan LN, Grow WB, Pusateri A, Douglas V, Mendenhall NP, 27 Carlotti E, Wrench D, Matthews J, et al. Transformation of follicular Lynch JW. Phase II trial of individualized rituximab dosing for lymphoma to diff use large B-cell lymphoma may occur by divergent patients with CD20-positive lymphoproliferative disorders. evolution from a common progenitor cell or by direct evolution J Clin Oncol 2005; 23: 1096–102. from the follicular lymphoma clone. Blood 2009; 113: 3553–57. 15 Kahl BS, Williams ME, Hong F, Gascoyne R, Horning SJ. 28 Ruminy P, Jardin F, Picquenot JM, et al. S(mu) mutation patterns Preliminary pharmacokinetic (PK) analysis of Eastern Cooperative suggest diff erent progression pathways in follicular lymphoma: Oncology Group Protocol E4402: Rituximab extended schedule or early direct or late from FL progenitor cells. Blood 2008; re-treatment trial (RESORT). Blood 2007; 110: 3420 (abstr). 112: 1951–59. 16 Cheson BD, Horning SJ, Coiffi er B, et al. Report of an international 29 Friedberg JW, Cohen P, Chen L, et al. Bendamustine in patients workshop to standardize response criteria for non-Hodgkin’s with rituximab-refractory indolent and transformed non-Hodgkin’s lymphomas. NCI Sponsored International Working Group. lymphoma: results from a phase II multicenter, single-agent study. J Clin Oncol 1999; 17: 1244–53. J Clin Oncol 2008; 26: 204–10. 17 Marcus R, Imrie K, Solal-Céligny P, et al. Phase III study of R-CVP 30 Kahl BS, Bartlett NL, Leonard JP, et al. Bendamustine is eff ective compared with cyclophosphamide, vincristine, and prednisone therapy in patients with rituximab-refractory, indolent B-cell alone in patients with previously untreated advanced follicular non-Hodgkin lymphoma: results from a Multicenter Study. lymphoma. J Clin Oncol 2008; 26: 4579–86. Cancer 2010; 116: 106–14. 18 Buske C, Hoster E, Dreyling M, et al. Rituximab in combination 31 Sebban C, Brice P, Delarue R, et al. Impact of rituximab and/or with CHOP in patients with follicular lymphoma: Analysis of high-dose therapy with autotransplant at time of relapse in patients treatment outcome of 552 patients treated in a randomized trial of with follicular lymphoma: a GELA study. J Clin Oncol 2008; the German Low Grade Lymphoma Study Group (GLSG) after a 26: 3614–20. follow up of 58 months. Blood 2008; 112: 2599 (abstr). 32 Bachy E, Brice P, Delarue R, et al. Long-term follow-up of patients 19 Solal-Céligny P, Roy P, Colombat P, et al. Follicular lymphoma with newly diagnosed follicular lymphoma in the prerituximab international prognostic index. Blood 2004; 104: 1258–65. era: eff ect of response quality on survival-A study from the 20 Common Terminology Criteria for Adverse Events v3.0 (CTCAE). Groupe d’Etude des Lymphomes de l’Adulte. J Clin Oncol 2010; August, 2006. http://ctep.cancer.gov/protocoldevelopment/ 28: 822–29. electronic_applications/docs/ctcaev3.pdf (accessed Nov 1, 2010). 33 Martinelli G, Schmitz SF, Utiger U, et al. Long-term follow-up of 21 Vidal L, Gafter-Gvili A, Leibovici L, et al. Rituximab maintenance for patients with follicular lymphoma receiving single-agent rituximab the treatment of patients with follicular lymphoma: systematic review at two diff erent schedules in trial SAKK 35/98. J Clin Oncol 2010; and meta-analysis of randomized trials. J Natl Cancer Inst 2009; 28: 4480–84. 101: 248–55.

www.thelancet.com Vol 377 January 1, 2011 51